NASA AI predicts solar active regions 12 hours early

NASA’s COFFIES Science Center has developed a machine-learning model that can predict the emergence of active regions on the Sun up to 12 hours before they appear. The model, described in the Journal of Geophysical Research: Machine Learning and Computation, uses a sliding-window transformer architecture to detect subtle changes in acoustic waves and magnetic fields beneath the solar surface.

The COFFIES team includes researchers from New Jersey Institute of Technology, Princeton University, and NASA’s Ames Research Center. They analyzed data from NASA’s Solar Dynamics Observatory and used supercomputing resources at Ames. Alexander Kosovichev, a COFFIES co-investigator at NJIT, said the technique identifies precursors in slight changes of the Sun’s acoustic power, comparing it to “a slight change in rhythm within a very noisy orchestra.”

Active regions are concentrations of magnetic fields that break through the solar surface as sunspots. They are the main engines behind solar flares and coronal mass ejections, which can threaten astronauts, disable satellites, and disrupt radio communications on Earth. Current operational forecasts from NOAA’s Space Weather Prediction Center and the U.S. Air Force monitor active regions already visible on the Sun to estimate flare probabilities.

The new AI model moves a fixed-size viewing window across a long timeline of solar activity, focusing on recent data while remembering overall patterns. This allows forecasters to predict approximate locations of emerging sunspots rather than relying on counting already visible sunspots. The model is not yet ready for operational real-time forecasting, but the team plans to validate it across many more known solar events.

COFFIES, which stands for Consequence Of Fields and Flows in the Interior and Exterior of the Sun, is one of three NASA DRIVE Science Centers. It focuses on interconnected processes behind the Sun’s activity and its 11-year cycle. Michelangelo Romano, deputy director of NASA’s Moon to Mars Space Weather Analysis Office, said the model could provide new capabilities for predicting potential flaring locations ahead of time, offering additional support to NASA missions.

NASA’s space weather monitoring is critical for the Artemis missions to the Moon and future crewed missions to Mars. Teams across NASA and NOAA collaborate to transition research into operational tools, including the Space Radiation Analysis Group, Moon to Mars Space Weather Analysis Office, and Community Coordinated Modeling Center. Sunspot region emergence prediction, especially for the Sun’s far side, could supplement current models used by these teams.